Light distillation for Incremental Graph Convolution Collaborative Filtering
X Fan, F Mo, C Chen, H Yamana

TL;DR
This paper introduces a lightweight distillation approach for incremental graph convolutional collaborative filtering, significantly reducing training time while maintaining or improving recommendation accuracy.
Contribution
It proposes a preference-driven distillation method that simplifies incremental GCN training, addressing the high complexity and time consumption of existing methods.
Findings
Reduces training time by up to 9.5x compared to existing methods.
Improves Recall@20 by over 5% in experiments.
Maintains performance with less complex model architecture.
Abstract
Recommender systems presently utilize vast amounts of data and play a pivotal role in enhancing user experiences. Graph Convolution Networks (GCNs) have surfaced as highly efficient models within the realm of recommender systems due to their ability to capture extensive relational information. The continuously expanding volume of data may render the training of GCNs excessively costly. To tackle this problem, incrementally training GCNs as new data blocks come in has become a vital research direction. Knowledge distillation techniques have been explored as a general paradigm to train GCNs incrementally and alleviate the catastrophic forgetting problem that typically occurs in incremental settings. However, we argue that current methods based on knowledge distillation introduce additional parameters and have a high model complexity, which results in unrealistic training time consumption…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsDigital Marketing and Social Media · Human Mobility and Location-Based Analysis · Technology Adoption and User Behaviour
MethodsConvolution · Knowledge Distillation
